真武汤治疗慢性心力衰竭的证据图谱。
全文（开放获取） · 中文译文
摘要 目的：尽管有指南指导的治疗，慢性心力衰竭（CHF）仍然与症状负担、反复住院和生活质量降低相关。真武汤（ZWD）是一种经典的中药方剂，用作阳虚水泛证的辅助治疗。本证据图谱旨在描述真武汤治疗慢性心力衰竭临床研究的数量、分布、方法学质量和证据空白。 方法：我们检索了CNKI、VIP、万方、CMJ、SinoMed、PubMed、Web of Science、Embase、Cochrane图书馆、ClinicalTrials.gov、ITMCTR和ChiCTR，检索时间从建库至2025年3月2日。评估原方或加减真武汤治疗CHF的随机对照试验（RCT）、系统评价和meta分析符合纳入资格。为检索敏感性使用了宽泛的心力衰竭术语，但最终资格限定为CHF。随机对照试验采用Cochrane偏倚风险评估工具评估，系统评价/meta分析采用AMSTAR-2和PRISMA 2020评估。 结果：排除一项急性心力衰竭研究后，修订后的证据基础包括166项随机对照试验和14项系统评价/Meta分析。所有纳入的临床研究均在中国进行，且通常为单中心、小样本、短疗程试验，随访有限。常见报告结局包括左心室射血分数、纽约心脏协会心功能分级和利钠肽水平。最常报告的中医证型为心肾阳虚和阳虚水泛。重要局限性包括随机化、分配隐藏、盲法、方案注册、随访和不良事件的报告不完整。所有纳入的系统评价/Meta分析经AMSTAR-2评定为极低质量。 结论：当前证据提示真武汤（ZWD）对慢性心力衰竭（CHF）可能存在辅助治疗信号，但证据确定性仍较低。未来研究应采用前瞻性注册、效能充足、多中心随机对照试验，并具有标准化辨证、较长随访、临床有意义的终点和严格的安全性监测。 引言 心力衰竭（HF）是一种由心脏结构或功能异常导致心室充盈或射血受损的临床综合征。其常见特征为呼吸困难、疲劳、运动不耐受和液体潴留。随着中国人口老龄化和心血管危险因素谱的变化，心力衰竭已成为住院、死亡、生活质量下降和社会经济负担的重要原因（–）。 指南指导的药物治疗可改善症状和结局，但许多慢性心力衰竭（CHF）患者仍会出现症状不稳定、反复失代偿和日常功能受损。在中国，中医（TCM）通常作为常规医疗的辅助手段使用。常规治疗根据心力衰竭表型选择，旨在改变已确立的、对发病率和死亡率有循证影响的心血管和心肾通路；而中医药治疗通常根据辨证论治个体化实施，并与标准治疗联用，而非替代标准治疗。对于心血管疾病，既往关于中草药制剂作为辅助治疗的证据综合报告了功能或替代结局改善的信号，部分分析中显示出较少的心血管事件或心力衰竭再住院；然而，方法学局限性、异质性和安全性与长期终点报告不足，限制了对指南指导药物治疗的直接比较（, , , , ）。中药复方可能通过涉及体液代谢、神经激素激活、炎症、氧化应激、微循环和心肌能量代谢的多组分作用影响症状负担和功能状态；然而，临床证据的确定性不一，需要进行批判性评价（–）。 真武汤（ZWD）源自《伤寒杂病论》，由附子、茯苓、白术、白芍和生姜组成。在中医理论中，用于温阳、利水、健脾、化饮。此处，心肾阳虚是指一种中医证型，其中心肾温煦和气化功能减退与乏力、畏寒、运动耐量下降及水液代谢障碍等表现相关。阳虚水泛是指一种相关证型，其特征为水液气化失常，伴有水肿、体重增加或其他水液潴留征象。这些概念用于中医辨证分型和治疗选择；它们不等同于西医心力衰竭表型或独立的生物医学诊断。其临床意义在于，真武汤传统上被选择用于具有临床可辨识的虚证和水液潴留表现群的患者，而诊断标准的差异可能引入试验间的异质性（, ）。因此，真武汤在慢性心力衰竭领域的临床研究日益受到关注。然而，该研究领域的分布、质量和局限性尚未得到系统性的梳理。 证据图谱是一种广泛的证据综合方法，用于描述某一领域研究的数量、特征、分布和空白。与传统的系统评价或荟萃分析不同，证据图谱的主要目的不是产生单一的合并效应估计。相反，它以视觉和叙述的方式总结证据集中在哪里、证据薄弱之处，以及哪些问题需要更严谨的研究 (, )。这种方法适用于充血性心力衰竭（CHF）中的真武汤（ZWD），因为文献在研究设计、方剂加减、治疗持续时间、结局、辨证和综述质量方面存在异质性。 材料与方法 文献检索 数据检索自中国知网（CNKI）、万方数据知识服务平台、维普中文科技期刊数据库、SinoMed、PubMed、Web of Science、Embase、Cochrane图书馆、中文医学期刊全文数据库（CMJ）、中国临床试验注册中心（ChiCTR）、国际传统医学临床试验注册平台（ITMCTR）和ClinicalTrials.gov，检索时间范围从建库至2025年3月2日。中文检索词包括“真武汤”、“心力衰竭”、“慢性心力衰竭”、“心衰”、“心脏失代偿”、“心功能不全”和“心衰竭”。英文检索词包括“Zhenwu decoction”、“Zhenwu Tang”、“heart failure”、“chronic heart failure”、“cardiac failure”、“cardiac insufficiency”、“heart decompensation”、“myocardial failure”和“congestive heart failure”。采用宽泛的心力衰竭（HF）检索词以提高检索敏感度，因为有些记录在标题或摘要中未标明是否为慢性，但最终符合纳入标准者仅限于慢性心力衰竭（CHF）。PubMed检索将MeSH主题词与标题/摘要检索词相结合，例如：(“Zhenwu decoction”[Title/Abstract] OR “Zhenwu Tang”[Title/Abstract] OR “Zhen Wu Tang”[Title/Abstract]) AND (“Heart Failure”[MeSH Terms] OR “heart failure”[Title/Abstract] OR “chronic heart failure”[Title/Abstract] OR “cardiac failure”[Title/Abstract] OR “cardiac insufficiency”[Title/Abstract] OR “congestive heart failure”[Title/Abstract])。检索式根据各个数据库进行了调整。学位论文、会议论文、试验注册记录及其他灰色文献来源，若符合纳入标准，均可纳入。文献检索与筛选均独立进行。 由两名研究者执行。 文献纳入标准 符合条件的研究是可公开获取的随机对照试验（RCT）、系统评价或荟萃分析，这些研究评价真武汤（ZWD）作为慢性心力衰竭（CHF）的干预措施。干预措施必须使用原方真武汤或保留附子、茯苓、白术、白芍和生姜核心成分的加减方。符合条件的对照措施包括标准西医疗法、安慰剂或标准治疗联合安慰剂。仅纳入中文或英文发表的文献。 文献排除标准 排除标准为：（1）数据相同的重复发表；（2）叙述性综述；（3）动物或体外研究；（4）可提取数据不足的研究；（5）无法获取全文；（6）真武汤仅给予对照组的研究；（7）当无法分离真武汤的独立效应时，试验组额外接受其他中医干预（如另一种草药方剂、中成药、针刺或艾灸）的试验；（8）RCT方案；（9）在核心方基础上增加或去除三种及以上草药组分的改良真武汤研究；以及（10）聚焦于急性心力衰竭而非慢性心力衰竭的研究。 文献筛选与数据提取 文献筛选和数据提取由两名研究者采用盲法独立进行。分歧通过讨论或咨询第三名评审者解决。参考文献使用EndNote 21管理。去重后，根据预定标准筛选标题和摘要，随后进行全文资格评估。提取项目包括：作者、年份、标题、研究设计、样本量、心力衰竭类型、中医证型、干预和对照方案、真武汤组成或化裁、治疗持续时间、随访、心功能分级、结局指标、总有效率、基础疾病、合并症以及不良事件报告。 采用Cochrane偏倚风险评估工具评估纳入的随机对照试验（RCT）的质量，系统评价/Meta分析采用AMSTAR-2评估。RCT偏倚风险评估在研究层面进行，因为本证据图谱旨在描述整体证据格局，而非生成针对特定结局的合并效应量。对于随机序列生成，仅在明确描述了充分的方法（如随机数字表或计算机生成的序列）时才判定为低偏倚风险；仅提及“随机”不视为充分。分配隐藏仅在报告了中心随机化、密封不透光信封或其他充分隐藏方法时才评为低风险。盲法、结局数据不完整、选择性报告和其他偏倚根据每篇已发表报告中可得的信息进行判断。 AMSTAR-2用于评价系统评价和Meta分析的方法学质量。关键领域为条目2、4、7、9、11、13和15。根据AMSTAR-2指南()，总体可信度评为高、中、低或极低。报告质量采用PRISMA 2020清单评估，每个条目根据报告完整性评为0、0.5或1分。 证据图谱与分类规则。 证据图用于总结发表趋势、干预措施特征、结局分布、中医证型、方法学质量、报告质量以及报告的效应方向。当一篇综述报告了至少两个预先指定的主要结局，或一个主要临床结局加上一致的支持性次要结局，显示有利于ZWD的统计学显著合并效应，且未报告严重不良事件增加时，判定为“有益”。当效应方向在至少一个主要结局上有利于ZWD，但确定性因异质性、不精确性、安全性报告稀疏、次要结局不一致、样本量有限或方法学质量极低而减弱时，判定为“潜在有益”。当合并结果未显示统计学显著或临床有意义的差异时，判定为“无差异”。当ZWD与更差的疗效结局或更高的严重不良事件或毒性发生率相关时，判定为“有害”。这些类别描述的是报告的效应方向，并不表示证据确定性高。 方案注册 本证据图未进行前瞻性注册。缺乏方案注册被认为是一个局限性，因为可能增加选择性方法学决定的风险。 结果 文献筛选结果 数据库和注册库检索识别出3,085条记录。去除1,834条重复记录后，1,251条记录通过标题和摘要进行筛选，排除389条。在寻求获取的862篇报告中，7篇不可获取。共855篇全文报告进行了合格性评估。排除的报告原因包括：干预措施不当（n = 603）、非合格的非随机对照试验设计（n = 54）、主题不相关（n = 17）、或聚焦于急性心力衰竭（n = 1）。最终证据库包含180项研究，包括166项随机对照试验和14项系统评价/Meta分析。图展示了PRISMA 2020流程图。 研究选择的PRISMA 2020流程图。 纳入随机对照试验的基线特征 修订后的证据库包括166项慢性心力衰竭随机对照试验，均在中国开展。对照组接受标准西医治疗或安慰剂，而治疗组在对照干预基础上接受原始或改良的参附汤（ZWD）。大多数试验为小型、单中心研究，干预持续时间短且随访有限。提取的研究层面变量包括样本量、慢性心力衰竭类型、干预方案、治疗持续时间、主要结局、随访和不良事件报告。 心力衰竭类型的统计学分析 在全文重新评估后，合格性限定为慢性心力衰竭。唯一一项关于急性心肌梗死后急性心力衰竭的研究被排除在修订后的证据库之外。其余随机对照试验评估的是慢性心力衰竭人群或被描述为慢性、稳定或长期疾病的心力衰竭人群。 103项研究报告了基线左心室射血分数（LVEF）值或LVEF亚型描述。原始提取将一些混合性或重叠性LVEF人群计入了多个亚型，导致亚型总数超过了报告LVEF的研究数量。在修订后的报告中，LVEF类别应作为互斥组报告：射血分数降低的心力衰竭（HFrEF）、射血分数轻度降低的心力衰竭（HFmrEF）、射血分数保留的心力衰竭（HFpEF）、混合/重叠性LVEF或LVEF不明确。因此，原始提取的亚型频率（HFrEF，n = 94；HFmrEF，n = 92；HFpEF，n = 20）报告为描述性报告频率，而非互斥的研究计数。 年度发文趋势 1999年至2025年间，ZWD治疗慢性心力衰竭（CHF）的年发文量呈波动上升趋势，2020年达到峰值18篇。2025年的数量仅包括截至2025年3月2日被索引的记录，不应解释为全年估计值（）。 ZWD治疗CHF的年发文量。 期刊发文量 纳入的研究发表于105种期刊、6篇硕士学位论文、2篇会议论文、1个研究网站出版物和1篇报纸文章。18种期刊各发表了至少3篇文章。28篇文章发表在核心期刊上，3篇被科学引文索引扩展版（SCIE）期刊收录。发文量排名前三的期刊是《中国中医药临床研究杂志》、《实用中医药杂志》和《中西医结合心脑血管病杂志》。 干预持续时间与随访期 十三项研究未报告干预持续时间。在其余试验中，多数干预持续15–30天。四项研究持续少于7天，24项持续7–15天，79项持续15天至1个月，30项持续5周至2个月，12项持续3个月，5项超过3个月。大多数试验未报告治疗后随访，限制了对持久性、再住院、死亡率和迟发性不良事件的评估。 证型分布分析 在166项慢性心力衰竭（CHF）随机对照试验（RCT）中，91项报告了受试者的中医证型，并描述了11种独特证型。心肾阳虚是最常见的证型（45/166，占所有CHF随机对照试验的27.1%；45/91，占报告证型的随机对照试验的49.5%），其次是阳虚水停或水泛（36/166，占所有CHF随机对照试验的21.7%；36/91，占报告证型的随机对照试验的39.6%）。其余证型报道较少，包括脾肾阳虚、气虚血瘀兼阳虚水泛、肾阳虚、肾阳衰微、痰饮阻肺、心肾阳虚兼阳虚水泛、阳虚水泛兼脾胃气虚、阳虚水泛兼血瘀，以及阳虚水停。 纳入研究的心功能分级 总共有118篇文章记录了心功能分级；其中87篇（73%）明确了每个分级的患者数。此外，在这87项研究中，37例被归为纽约心脏协会（NYHA）I级，2,555例为II级，3,494例为III级，1,570例为IV级。 纳入随机对照试验（RCT）中原发疾病的分布 结果还显示，71篇文章明确了入组患者的心力衰竭原发疾病，包括冠心病、高血压、肺源性心脏病、风湿性心脏病、扩张型心肌病、甲状腺功能亢进性心脏病、心肌病、心脏瓣膜病、心律失常、先天性心脏病、糖尿病和急性心肌梗死。其中排名前三的疾病为冠心病（n = 2,429）、高血压（n = 1,182）和肺源性心脏病（n = 854）。 纳入文献的质量评价 纳入的随机对照试验采用Cochrane偏倚风险工具进行评估（）。原始报告经常使用“随机化”一词，但许多报告未提供足够细节以验证序列生成或分配隐藏。参与者、研究人员和结局评估的盲法很少被报告。因此，整体随机对照试验证据基础存在重要的方法学和报告局限性，包括盲法不充分、分配隐藏不明确、方案可获性有限、随访时间短以及安全性报告不完整。偏倚风险结果应被解释为基于报告的判断，而非对试验高质量的确证。 纳入随机对照试验的偏倚风险评估。 系统评价的方法学质量评价 采用AMSTAR-2工具评估纳入的14项系统评价/Meta分析的方法学质量（–）。所有均被评为极低质量。常见的关键缺陷包括：缺少注册或预先指定的方案、检索报告不完整、缺乏列出排除研究及其理由的清单、对原始研究偏倚风险的处理不充分，以及发表偏倚评估不足（）。 纳入系统评价/Meta分析的AMSTAR-2条目级评估。 证据图谱 在14项系统评价/Meta分析中，根据预设规则，9项被归类为报告效应方向有益，5项被归类为报告潜在有益方向。没有系统评价被归类为显示无差异或有害效应。这些标签仅描述所报告的效应方向，不应被解释为高确定性。由于所有纳入的系统评价经AMSTAR-2评级均为极低质量，因此证据图谱中分别呈现效应方向和方法学质量（）。 ZWD治疗CHF的证据图谱。效应方向与AMSTAR-2方法学质量应分别解读。 系统评价报告的质量评估 报告质量
全文（开放获取） · 英文原文
【Abstract】 OBJECTIVE: Chronic heart failure (CHF) remains associated with symptom burden, recurrent hospitalization, and reduced quality of life despite guideline-directed therapy. Zhenwu Decoction (ZWD), a classical traditional Chinese medicine formula, is used as adjunctive therapy for yang-deficiency and fluid-retention patterns. This evidence map aimed to characterize the volume, distribution, methodological quality, and evidence gaps in clinical research on ZWD for CHF. METHODS: We searched CNKI, VIP, Wanfang, CMJ, SinoMed, PubMed, Web of Science, Embase, the Cochrane Library, ClinicalTrials.gov, ITMCTR, and ChiCTR from inception to 2 March 2025. Randomized controlled trials (RCTs), systematic reviews, and meta-analyses evaluating original or modified ZWD for CHF were eligible. Broad heart-failure terms were used for search sensitivity, but final eligibility was restricted to CHF. RCTs were assessed with the Cochrane risk-of-bias tool, and systematic reviews/meta-analyses were assessed with AMSTAR-2 and PRISMA 2020. RESULTS: After exclusion of one acute heart failure study, the revised evidence base comprised 166 RCTs and 14 systematic reviews/meta-analyses. All included clinical studies were conducted in China and were generally single-center, small-sample, and short-duration trials with limited follow-up. Commonly reported outcomes included left ventricular ejection fraction, New York Heart Association functional class, and natriuretic peptide levels. The most frequently reported traditional Chinese medicine syndrome patterns were heart-kidney yang deficiency and yang deficiency with water retention. Important limitations included incomplete reporting of randomization, allocation concealment, blinding, protocol registration, follow-up, and adverse events. All included systematic reviews/meta-analyses were rated as critically low quality by AMSTAR-2. CONCLUSION: Current evidence suggests potential adjunctive signals for ZWD in CHF, but the certainty of evidence remains low. Future studies should use prospectively registered, adequately powered, multicenter RCTs with standardized syndrome differentiation, longer follow-up, clinically meaningful endpoints, and rigorous safety monitoring. 【Introduction】 Heart failure (HF) is a clinical syndrome caused by structural or functional cardiac abnormalities that impair ventricular filling or ejection. It is commonly characterized by dyspnea, fatigue, exercise intolerance, and fluid retention. With population aging and changes in cardiovascular risk profiles in China, HF has become an important cause of hospitalization, mortality, reduced quality of life, and socioeconomic burden (–). Guideline-directed medical therapy can improve symptoms and outcomes, but many patients with chronic heart failure (CHF) continue to experience unstable symptoms, recurrent decompensation, and impaired daily function. In China, traditional Chinese medicine (TCM) is commonly used as an adjunct to conventional medical care. Conventional therapy is selected according to heart-failure phenotype and is intended to modify established cardiovascular and cardiorenal pathways with evidence-based effects on morbidity and mortality, whereas TCM treatment is generally individualized according to syndrome differentiation and used alongside, rather than instead of, standard therapy. For cardiovascular disease, previous evidence syntheses of Chinese herbal formulations used as adjuncts have reported signals for improvement in functional or surrogate outcomes and, in some analyses, fewer cardiovascular events or heart-failure rehospitalizations; however, methodological limitations, heterogeneity, and sparse safety and long-term endpoint reporting limit direct comparison with guideline-directed therapy (, , , , ). TCM formulas may influence symptom burden and functional status through multi-component actions related to fluid metabolism, neurohormonal activation, inflammation, oxidative stress, microvascular circulation, and myocardial energy metabolism; however, the certainty of clinical evidence varies and requires critical appraisal (–). Zhenwu Decoction (ZWD) is derived from the Treatise on Cold Damage and Miscellaneous Diseases and comprises Aconiti Lateralis Radix Praeparata, Poria, Atractylodis Macrocephalae Rhizoma, Paeoniae Radix Alba, and Zingiberis Rhizoma Recens. In TCM theory, it is used to warm yang, promote water metabolism, strengthen the spleen, and reduce fluid retention. Here, heart-kidney yang deficiency denotes a TCM pattern in which reduced warming and transformative functions attributed to the heart and kidney are associated with manifestations such as fatigue, cold intolerance, reduced exercise capacity, and impaired fluid handling. Yang deficiency with water retention denotes a related pattern in which impaired fluid transformation is accompanied by oedema, weight gain, or other signs of fluid accumulation. These constructs are used for TCM pattern differentiation and treatment selection; they are not equivalent to Western heart-failure phenotypes or standalone biomedical diagnoses. Their clinical implication is that ZWD is traditionally selected for a clinically recognizable cluster of deficiency and fluid-retention manifestations, while variation in diagnostic criteria can introduce heterogeneity across trials (, ). ZWD has therefore attracted increasing clinical research interest for CHF. However, the distribution, quality, and limitations of this research field have not been systematically mapped. Evidence mapping is a broad evidence-synthesis approach that describes the volume, characteristics, distribution, and gaps of research in a field. Unlike a conventional systematic review or meta-analysis, an evidence map does not primarily aim to generate a single pooled effect estimate. Instead, it visually and narratively summarizes where evidence is concentrated, where it is weak, and which questions require more rigorous study (, ). This approach is appropriate for ZWD in CHF because the literature is heterogeneous in study design, formula modification, treatment duration, outcomes, syndrome differentiation, and review quality. 【Materials and methods】 【Literature search】 The data were retrieved from China National Knowledge Infrastructure (CNKI), Wanfang Data Knowledge Service Platform, VIP Chinese Science and Technology Periodical Database, SinoMed, PubMed, Web of Science, Embase, the Cochrane Library, Chinese Medical Journals Full-text Database (CMJ), Chinese Clinical Trial Registry (ChiCTR), International Traditional Medicine Clinical Trial Registry Platform (ITMCTR), and ClinicalTrials.gov from inception to 2 March 2025. Chinese search terms included “真武汤”, “心力衰竭”, “慢性心力衰竭”, “心衰”, “心脏失代偿”, “心功能不全”, and “心衰竭”. English search terms included “Zhenwu decoction”, “Zhenwu Tang”, “heart failure”, “chronic heart failure”, “cardiac failure”, “cardiac insufficiency”, “heart decompensation”, “myocardial failure”, and “congestive heart failure”. Broad HF terms were used to maximize sensitivity because some records do not specify chronicity in the title or abstract, but final eligibility was restricted to CHF. The PubMed search combined MeSH and Title/Abstract terms, for example: (“Zhenwu decoction”[Title/Abstract] OR “Zhenwu Tang”[Title/Abstract] OR “Zhen Wu Tang”[Title/Abstract]) AND (“Heart Failure”[MeSH Terms] OR “heart failure”[Title/Abstract] OR “chronic heart failure”[Title/Abstract] OR “cardiac failure”[Title/Abstract] OR “cardiac insufficiency”[Title/Abstract] OR “congestive heart failure”[Title/Abstract]). Search expressions were adapted for each database. Dissertations, conference papers, trial registries, and other grey literature sources were eligible if they met the inclusion criteria. Literature search and screening were independently performed by two investigators. 【Literature inclusion criteria】 Eligible studies were publicly accessible RCTs, systematic reviews, or meta-analyses that evaluated ZWD as an intervention for CHF. The intervention had to use the original ZWD formula or a modified formula retaining the core components of Aconiti Lateralis Radix Praeparata, Poria, Atractylodis Macrocephalae Rhizoma, Paeoniae Radix Alba, and Zingiberis Rhizoma Recens. Eligible comparators included standard Western medical care, placebo, or standard care combined with placebo. Only Chinese- or English-language publications were included. 【Literature exclusion criteria】 Exclusion criteria were: (1) duplicate publications with identical data; (2) narrative reviews; (3) animal or in vitro studies; (4) studies with insufficient extractable data; (5) inaccessible full text; (6) studies in which ZWD was administered only to the control group; (7) trials in which the experimental group received additional TCM interventions, such as another herbal formula, Chinese patent medicine, acupuncture, or moxibustion, when the independent effect of ZWD could not be separated; (8) RCT protocols; (9) studies in which modified ZWD added or removed three or more herbal components from the core formula; and (10) studies focused on acute heart failure rather than CHF. 【Literature screening and data extraction】 Literature screening and data extraction were independently conducted by two investigators using a blinded approach. Disagreements were resolved through discussion or consultation with a third reviewer. References were managed with EndNote 21. After duplicate removal, titles and abstracts were screened according to predetermined criteria, followed by full-text eligibility assessment. Extracted items included author, year, title, study design, sample size, type of heart failure, TCM syndrome pattern, intervention and control regimens, ZWD composition or modification, treatment duration, follow-up, cardiac function classification, outcome measures, overall response rate, underlying disease, comorbidities, and adverse-event reporting. 【Quality assessment】 The quality of included RCTs was assessed with the Cochrane risk-of-bias tool, and systematic reviews/meta-analyses were assessed with AMSTAR-2. RCT risk-of-bias assessment was conducted at the study level because this evidence map aimed to describe the overall evidence landscape rather than produce outcome-specific pooled effects. For random sequence generation, low risk was assigned only when an adequate method, such as a random-number table or computer-generated sequence, was explicitly described. Merely stating “randomized” was not considered sufficient. Allocation concealment was rated as low risk only when central randomization, sealed opaque envelopes, or another adequate concealment method was reported. Blinding, incomplete outcome data, selective reporting, and other bias were judged according to the information available in each published report. AMSTAR-2 was used to evaluate the methodological quality of systematic reviews and meta-analyses. The critical domains were items 2, 4, 7, 9, 11, 13, and 15. Overall confidence was rated as high, moderate, low, or critically low according to AMSTAR-2 guidance (). Reporting quality was assessed using the PRISMA 2020 checklist, with each item scored as 0, 0.5, or 1 according to completeness of reporting. 【Evidence mapping and classification rules】 Evidence maps were used to summarize publication trends, intervention characteristics, outcome distribution, TCM syndrome patterns, methodological quality, reporting quality, and the reported direction of effects. “Beneficial” was assigned when a review reported statistically significant pooled effects favoring ZWD on at least two prespecified major outcomes, or on one major clinical outcome plus consistent supportive secondary outcomes, without a reported increase in serious adverse events. “Potentially beneficial” was assigned when the direction of effect favored ZWD for at least one major outcome but certainty was weakened by heterogeneity, imprecision, sparse safety reporting, inconsistent secondary outcomes, limited sample size, or critically low methodological quality. “No difference” was assigned when pooled results did not show a statistically significant or clinically meaningful difference. “Harmful” was assigned when ZWD was associated with worse efficacy outcomes or a higher rate of serious adverse events or toxicity. These categories describe reported direction of effect and do not indicate high certainty of evidence. 【Protocol registration】 This evidence map was not prospectively registered. The absence of protocol registration is acknowledged as a limitation because it may increase the risk of selective methodological decisions. 【Results】 【Literature selection outcomes】 The database and registry search identified 3,085 records. After removal of 1,834 duplicates, 1,251 records were screened by title and abstract, and 389 were excluded. Of 862 reports sought for retrieval, 7 were unavailable. A total of 855 full-text reports were assessed for eligibility. Reports were excluded because of inappropriate intervention (n = 603), non-eligible non-RCT design (n = 54), inappropriate topic (n = 17), or focus on acute heart failure (n = 1). The final evidence base comprised 180 studies, including 166 RCTs and 14 systematic reviews/meta-analyses. illustrates the PRISMA 2020 flow diagram. PRISMA 2020 flow diagram of study selection. 【Baseline characteristics of included RCTs】 The revised evidence base included 166 CHF RCTs, all conducted in China. The control groups received standard Western medical care or placebo, whereas the treatment groups received original or modified ZWD in addition to the control intervention. Most trials were small, single-center studies with short intervention durations and limited follow-up. Extracted study-level variables included sample size, CHF type, intervention regimen, treatment duration, main outcomes, follow-up, and adverse-event reporting. 【Statistical analysis of heart failure types】 After full-text reassessment, eligibility was restricted to CHF. The single study focused on acute heart failure after acute myocardial infarction was excluded from the revised evidence base. The remaining RCTs evaluated CHF populations or heart failure populations described as chronic, stable, or long-term disease. Baseline LVEF values or LVEF subtype descriptions were reported in 103 studies. The original extraction counted some mixed or overlapping LVEF populations in more than one subtype, which caused subtype totals to exceed the number of studies reporting LVEF. In the revised presentation, LVEF categories should be reported as mutually exclusive groups: HFrEF, HFmrEF, HFpEF, mixed/overlapping LVEF, or unclear LVEF. The originally extracted subtype frequencies (HFrEF, n = 94; HFmrEF, n = 92; HFpEF, n = 20) are therefore reported as descriptive reporting frequencies rather than mutually exclusive study counts. 【Annual publication trend】 Between 1999 and 2025, the annual number of publications on ZWD for CHF fluctuated upward and peaked at 18 papers in 2020. The 2025 count includes records indexed only up to 2 March 2025 and should not be interpreted as a full-year estimate (). Annual publications on ZWD for CHF. 【Journal publication volume】 The included studies were published in 105 journals, 6 master's theses, 2 conference papers, 1 research website publication, and 1 newspaper article. Eighteen journals published at least three articles each. Twenty-eight articles were published in core journals, and three were indexed in SCIE journals. The top three journals by publication volume were Chinese Journal of Clinical Research on Traditional Chinese Medicine, Journal of Practical Traditional Chinese Medicine, and Chinese Journal of Integrative Medicine on Cardio/Cerebrovascular Disease. 【Intervention duration and follow-up period】 Thirteen studies did not report intervention duration. Among the remaining trials, most interventions lasted 15–30 days. Four studies lasted less than 7 days, 24 lasted 7–15 days, 79 lasted 15 days to 1 month, 30 lasted 5 weeks to 2 months, 12 lasted 3 months, and 5 exceeded 3 months. Most trials did not report post-treatment follow-up, limiting assessment of durability, rehospitalization, mortality, and delayed adverse events. 【Pattern distribution analysis】 Among the 166 CHF RCTs, 91 reported participants’ TCM syndrome patterns and described 11 unique patterns. Heart-kidney yang deficiency was the most frequent pattern (45/166, 27.1% of all CHF RCTs; 45/91, 49.5% of RCTs reporting syndrome patterns), followed by yang deficiency with water retention or water flooding (36/166, 21.7% of all CHF RCTs; 36/91, 39.6% of RCTs reporting syndrome patterns). The remaining patterns were reported infrequently and included spleen-kidney yang deficiency, qi deficiency with blood stasis and yang deficiency accompanied by water flooding, kidney yang deficiency, kidney yang decline, phlegm-fluid obstructing the lung, heart-kidney yang deficiency with yang deficiency and water flooding, yang deficiency with water flooding and spleen-stomach qi deficiency, yang deficiency with water flooding and blood stasis, and yang deficiency with water retention. 【Heart function classification in included studies】 In total, 118 articles documented the heart function classification; of these, 87 papers (73%) specified the number of patients in each category. Furthermore, among the 87 studies, 37 cases were categorized as NYHA Class I, 2,555 as Class II, 3,494 as Class III, and 1,570 as Class IV. 【Distribution of primary diseases in included RCTs】 The results also showed that 71 articles specified the primary heart failure diseases of enrolled patients, including coronary heart disease, hypertension, pulmonary heart disease, rheumatic heart disease, dilated cardiomyopathy, hyperthyroid heart disease, cardiomyopathy, valvular heart disease, arrhythmia, congenital heart disease, diabetes, and acute myocardial infarction. The top three of these diseases were coronary heart disease (n = 2,429), hypertension (n = 1,182), and pulmonary heart disease (n = 854). 【Quality assessment of included literature】 The included RCTs were assessed with the Cochrane risk-of-bias tool (). The original reports frequently used the term “randomized”, but many did not provide enough detail to verify sequence generation or allocation concealment. Blinding of participants, personnel, and outcome assessment was rarely reported. The overall RCT evidence base therefore has important methodological and reporting limitations, including inadequate blinding, unclear allocation concealment, limited protocol availability, short follow-up, and incomplete safety reporting. The risk-of-bias results should be interpreted as reporting-based judgments rather than confirmation of high trial quality. Risk-of-bias assessment of included RCTs. 【Methodological quality assessment of systematic reviews】 The AMSTAR-2 tool was used to evaluate the methodological quality of the 14 included systematic reviews/meta-analyses (–). All were rated as critically low quality. Common critical weaknesses included absence of a registered or prespecified protocol, incomplete search reporting, lack of a list of excluded studies with reasons, inadequate handling of risk of bias in primary studies, and insufficient assessment of publication bias (). AMSTAR-2 item-level assessment of included systematic reviews/meta-analyses. 【Evidence map】 Among the 14 systematic reviews/meta-analyses, 9 were classified as reporting a beneficial direction of effect and 5 as reporting a potentially beneficial direction of effect according to the predefined rules. No review was classified as showing no difference or harmful effects. These labels describe the reported direction of effect only and should not be interpreted as high certainty. Because all included reviews were rated as critically low quality by AMSTAR-2, the direction of effect and methodological quality are presented separately in the evidence map (). Evidence map of ZWD for CHF. Direction of effect and AMSTAR-2 methodological quality should be interpreted separately. 【Quality assessment of systematic review reporting】 Reporting quality